> How is that different to the current VS-DR?
I suppose it's not except for the addition of a dynamic routing protocol
that determines which director is available.
Furthermore, you could potentially use the routing protocol to handle load
balancing of the directors, so you can set up an active/active/active
situation (N+1) for more scalability, rather than simple failover. Of
course this is contingent upon being able to cause the router to remember
previous gateways, or upon LVS propogating state information to other
directors, but I think the first can be done, and I know the second is
being worked on.
Mind you, all of this could very well be done by other methods, but this
one is ready and available, and you can use proven protocols to manage it.
Assuming it works. :)
Thanks,
Kyle Sparger
|